1

I am using SAPI 5.4 and wonder if it is possible to make the voice regonization better understand me i general or just to match my voice against certain phrases?

for example by recording certain phrases and match it against a phrase or so.

Even if it is possible to make it recognize if it is me who is speaking or someone else.

//Trind

Trind
  • 1,583
  • 4
  • 18
  • 38
  • 1
    Well that is for the speach recognition in windows not when coding, and the recognizer understands me but i want it but learn better how i am pronouncing, and i am not using the System.speech. – Trind Jun 11 '13 at 12:30
  • I don't understand your question, then. Are you trying to train the recognizer in general, in which case the Microsoft guide above will guide you through the training process, or are you trying to make the [recognizer help you write code](http://stackoverflow.com/questions/1408874/speech-recognition-programming/1682389#1682389)? If the latter, you're pretty much out of luck. – Eric Brown Jun 12 '13 at 17:38
  • 1
    No i don't want the recognizer to write code, i want to train my speech engine, update its dictation library, but the link you gave to me is about the Client version of microsoft speech, i am talking about the service side one, the one that uses the Microsoft.Speech namespace. – Trind Jun 12 '13 at 18:41

2 Answers2

2

The server engine from Microsoft doesn't support adaptation. If you want to train your engine, try other more flexible engine like CMUSphinx, see the acoustic model adaptation tutorial for example

Nikolay Shmyrev
  • 24,897
  • 5
  • 43
  • 87
0

The server engine (in the Microsoft.Speech namespace) doesn't support dictation. You can't do training using the C# SAPI API (either System.Speech or Microsoft.Speech); you would need to use the native C++ APIs to do training. (That being said, I'm not 100% sure the server engine supports training; it's designed to be more or less speaker-independent.)

Community
  • 1
  • 1
Eric Brown
  • 13,774
  • 7
  • 30
  • 71